Micromorphology of Paleosols of the Marília Formation and their Significance in the Paleoenvironmental Evolution of the Bauru Basin, Upper Cretaceous, Southeastern Brazil

Deduction of associated paleoenvironments and paleoclimate, definition of the chronosequence of paleosols, and paleogeographic reconstruction have become possible through the application of micromorphology in paleopedology. Micromorphology has also been useful in recognition of weathering processes and definition of minerals formed in succession. In this respect, the objective of this study was to identify the development of pedogenic processes and discuss their significance in the paleoclimate evolution of the Marília Formation (Maastrichtian) of Bauru Basin. Three sections of the Marília Formation (A1, A2, and A3) were described, comprising nine profiles. Micromorphologic al analysis was carried out according to the specialized literature. In the Marília Formation, the paleosols developed in sandstones have argillic (Btkm, Bt) and carbonate (Bk) horizons with different degrees of cementation, forming mainly calcretes. The evolution of pedogenic processes, in light of micromorphological analysis, evidenced three moments or stages for the genesis of paleosols with Bkm, Btk, and Bt horizons, respectively. In the Maastrichtian in the Bauru Basin, the paleosols with Bkm are older and more arid environments, and those with Bt were formed in wetter weather, but not enough to lead to the genesis of enaulic-related distributions, typical of current Oxisols.


INTRODUCTION
Micromorphology or micropedology can be defined as the branch of soil science that deals with the description, interpretation and, to some extent, the measure of soil constituents, features, and fabrics at the microscopic level (Bullock et al., 1985). Among its main objectives, micromorphological analysis allows us to formulate hypotheses or statements about the genetic and evolutionary dynamics of soil in an attempt to clarify controversy regarding its origin, evolution, and behavior (Castro et al., 2003). The identification of different constituents of the soil in different fractions, as well as definition of possible interrelationships between them, is another key objective of micromorphology (Castro et al., 2003).
One of the applications of micromorphology is to identify the pedogenetic process and display transfers and concentrations of elements (eluviation, illuviation, nodulation, or concretion) and to follow the development of distinctive features (porosity and pedoturbation, among others). It can also be applied in recognition of the weathering process involved and the mineral succession formed (Delvigne, 1998). The transformation of saprolite to soil, called pedoplasmation, is another topic that can be studied by micromorphology, as it concerns in the first place the transformation of fabrics, initially without mineralogical or chemical changes (Stoops, 2008).
In paleopedology, micromorphology may be useful in deduction of paleoenvironments and associated paleoclimates, the chronosequence of paleosols, and paleogeographic reconstitution (Porta et al., 1999).
Paleosols are defined as soils that were formed on an ancient landscape (Wright, 1986). They may be buried soils and/or soils incorporated into the sedimentary sequences, exhumed soils, or soils developed on ancient relief surfaces (relict soils) (Andreis, 1981;Catt, 1990). Exposed on the surface and influenced by later environmental changes (Retallack, 2001), paleosols reveal ancient environments and contain records regarding the climate, vegetation, landforms, intensity of pedogenesis, and sedimentation rates in effect during their training (Kraus, 1992;Wright, 1992).
Paleosols have been used both in analysis and paleoenvironmental reconstructions (Andreis, 1981;Catt, 1990;Retallack, 2001;Sheldon and Tabor, 2009;Nascimento et al., 2017) by forming themselves into open systems capable of recording environmental conditions during their training, as in stratigraphic correlation (Brown and Kraus, 1988;Wright, 1992;Marriott and Wright, 1993;Kraus 1997;McCarthy and Plint, 2003), indicating a stable surface and representing moments of pause in erosion and deposition.
However, paleosols are often difficult to identify since various physical and chemical properties change as a result of weathering processes and infiltration of solutions. However, micromorphological features are well preserved and allow determination of the type of soil, the environment, and the associated paleoclimate (Stoops, 2008). Thus, micromorphology has become an important tool in studies on global climate change.
In the Marília Formation, the paleosols developed on sandstones have argillic (Btk, Bt) and calcic (Bk) horizons with different degrees of cementation, forming mainly calcretes.
Groundmass and coating features, hypocoatings, quasicoatings, and infillings reveal the pedological nature of calcretes as a result of in situ weathering, allowing determination of chronology, intensity of processes, and the climatic conditions of their occurrence.
Studies of micromorphological features reveal pedogenic processes and assist in the interpretation of paleoclimatic environments.
Thus, the objective of this study was to identify, through micromorphology, the development of pedogenic processes in paleosols of the Marília Formation and discuss their significance in paleoclimatic evolution of the Maastrichtian of the Bauru Basin, southeastern Brazil.

Characterization of the Bauru Basin and study area
The Bauru Basin, located in southeastern Brazil, covers an area of approximately 330,000 km², including the middle west of São Paulo, northeastern Mato Grosso do Sul, southern Mato Grosso, southern Goiás, and western Minas Gerais (Figure 1). This basin has an elliptical shape, elongated in the N-NE direction, and consists primarily of siliciclastic continental deposits (Batezelli, 2003).
The erosional processes responsible for the current configuration of the Bauru Basin boundaries are relates to tectonic restructuring of the Tertiary. This event is marked on the eastern border of the Bauru Basin by the Serra do Mar mountain range and Alto Paranaíba Uplift (Batezelli, 2003(Batezelli, , 2015.
The Bauru Basin originated in the Late Cretaceous, was developed over the basalt rocks of the Serra Geral Formation, and was generated by flexural subsidence caused by the weight of thick basalt and by Alto Paranaíba Uplift and Alkaline Province of Goiás ( Figure 1) (Riccomini, 1995(Riccomini, , 1997Fernandes and Coimbra, 2000;Batezelli 2003;Batezelli et al., 2007;Batezelli, 2010Batezelli, , 2015. The Bauru Basin developed in the Upper Cretaceous in a period after breakup of the Gondwana continent, and completion occurred in a climate with variations from semi-arid to arid between the Campanian (83.6 to 72.1 Ma B.P) and Maastrichtian (72.1 to 66 Ma B.P.) (Batezelli, 2015).
By analysis of facies, architectural elements, and Paleocurrent, Batezelli et al. (2007) concluded that the deposits of the Bauru Group were formed from unidirectional and gravitational flows of high energy, associated with proximal and intermediate portions of alluvial system dominated by braided rivers (Stanistreet and McCarthy, 1993) or distributive fluvial systems (Hartley et al., 2010) arising from the Alto Paranaíba Uplift and the Alkaline Province of Goiás ( Figure 1).
The Bauru Basin is divided into two groups: the Caiuá Group and Bauru Group (Figures 1 and 2). However, there are two different views on the lithostratigraphic position of these two groups. Authors such as Coimbra (1996, 2000) and Fernandes (2004) argue that the two groups are contemporary. Other authors (Fulfaro and Perinotto, 1996;Paula and Silva et al., 2005;Batezelli, 2010Batezelli, , 2015 put the Caiuá Group in the lower portion of the basin, separated from the Bauru Group with a stratigraphic unconformity, signaled by a very evolved paleosol (Geossol Santo Anastásio) highlighted by Fulfaro et al. (1999). Recent studies by Batezelli et al. (2007) and Batezelli (2010Batezelli ( , 2015 showed that the two groups are not contemporary (Figure 2), resolving disputes on the stratigraphy of the basin.
The Bauru Group in the state of São Paulo consists of Araçatuba, Adamantina (Vale do Rio do Peixe according to Fernandes and Coimbra, 2000) and Marília (Echaporã Member) formations, from base to top (Batezelli, 2003(Batezelli, , 2010(Batezelli, , 2015. For Batezelli (2003Batezelli ( , 2010Batezelli ( , 2015 the Araçatuba Formation was formed in a lacustrine environment (playa-lake) thatserved asthe base level for the river system (generator of the Adamantina and Marília formations). Filling occurred because of the progradational advance of an alluvial system dominated by a braided river that gave rise to the Marília Formation.
Sedimentary evolution of the alluvial system was marked by periods of fluvial sedimentation and aeolian reworking, interspersed with periods of non-deposition (Batezelli, 2010(Batezelli, , 2015. During times without deposition, the floodplain would be covered by vegetation and soil could develop. Thus, the Marília Formation consists of a succession of deposits and paleosols that record sedimentation and pedogenesis during the Maastrichtian of the Bauru Basin. In the Marília Formation, paleosols are constituted by argillic horizons (Btk and Bt) and a calcic horizon (Bk), with different degrees of cementation, constituting calcretes. The irregular distribution and different thicknesses of the profiles are related to the type of the parent material, hydrology, topography, and biology, as well as the exposure time of deposits to weathering agents.

Description of paleosols and sampling
Three sections (A1, A2, and A3) of the Marília Formation ( Figure 1) were described, and nine sample profiles collected. Characterization of paleosols was performed according to Retallack (2001)  In characterizing the field, ped structure, horizon, and root marks were identified, the three main attributes in recognition of paleosols (Andreis, 1981;Retallack, 1988;Catt, 1990;Retallack, 2001). To define the horizons, the following types of properties were used: texture, structure, color, thickness, and depth of horizons; types of contacts between paleosol horizons and sediment; presence, type, and size of aggregates (peds); presence and type of films (cutans), presence and type of nodules or cementation; bioturbations; presence or absence of mottling; type, shape, and percentage of root marks; and presence or absence of gleization and friction surfaces (slickensides). Calcic (Bkm) and argillic (Btkm and Bt) horizons have been identified in paleosols of the study area.

Micromorphological analysis
Description of the oriented thin sections was made according to Bullock et al. (1985) and Stoops (2003) using a binocular magnifying glass to separate domains on the thin section, and a petrographic microscope with magnification of 2.5X to 40X for analysis of the groundmass (s-matrix) and pedofeatures. The grain size class was determined indirectly by optical microscopy, as proposed by Fitzpatrick (1984) and Bullock et al. (1985).
Scanning electron microscopy analyses were performed on four samples, representative of the different horizons of paleosols.
Interpretations of the groundmass and pedofeatures were based on the studies of Delvigne (1998) and .
The Visilog 5.4 and Image J 1.48 software was used for processing the images.

Morphological characterization of paleosols
The synthesis of the micromorphological properties of paleosol profiles are summarized in tables 1, 2, and 3.
The identification of pedogenic structures (blocky, prismatic, and laminar) and mark roots, associated with the absence of stratifications, predominance of massive structures, and discontinuity of carbonate cementation on the basis of outcrops, were the main factors used to define the profiles, as being predominantly pedologic. There was a predominance of nodular calcretes, at the expense of powdery (friable), laminar, and hard crust calcretes.
In morphological characterization, paleosols were identified with a Bkm horizon and structure blocks in the Botucatu section (A1), argillic horizons (Btkm and Bt) with prismatic and blocky structure in the Piratininga section (A2), and Btkm and Bkm horizons in the Garça section (A3), with laminar, prismatic, and blocky structures ( Figure 3).

Micromorphology
Micromorphology revealed the predominance of porphyric c/f-related distribution in three sections, with some chitonic, enaulic, and gefuric regions.
The paleosols with Bkm horizons showed open porphyric-related distribution with chitonic regions and very few gefuric regions (Table 4, Figure 4a).
Close porphyric-related distribution with chitonic regions predominated in paleosols with Btkm and Bt horizons (Table 4; Figures 4c and 4d).

Interpretation of micromorphology
The way the fine material, the coarse material, and the voids are distributed provides clues regarding the origin and evolution of calcretes.
Interpretation of secondary processes of carbonate accumulation, such as substitution and recrystallization (Figures 5a, 6a and   The microcodium structures consist of cell aggregates composed of individual crystals of calcite (Kosir, 2004). Generally, microcodium is associated with the rhizogenic horizon ( Figure 7c). Plant roots and microorganisms associated with the rhizosphere produce significant accumulations of calcium carbonate near the surface (Kosir, 2004). Microcodium occurs in petrocalcic horizons and calcretes, unlike rhizoliths, which normally occur in soil and sediments not completely cemented by calcium carbonate (Durand et al., 2010). This feature is associated with Cretaceous paleosol flood plains, lacustrine deposits, and paludal deposits (Durand et al., 2010). Microcodium structures (Figure 7c) are evidence of calcretes of pedogenic origin and indicate biologically controlled precipitation of calcium carbonate. Accumulations of microcodium probably reflect specific types of vascular plants of a pioneer community that had the ability to colonize carbonate substrates during the early phases of subaerial exposure (Kosir, 2004).
Groundmass features and pedofeatures have shown the predominance of beta type microfabrics in calcretes of the Marília Formation. Beta type microfabrics (biogenic/microbial related to the presence of roots, carbonates with alveolar fabric, fibrous calcite, Microcodium) and the presence of meniscus structures and pending cementation (pendant) are typical and striking features in pedogenic calcretes (Pimentel et al., 1996).
The calcite pendants (pending calcite) were another striking features in calcretes of the Marília Formation, manifesting their pedogenic origins. The pendants are laminated and composed of micritic and microsparitic calcite, with color ranging from light to dark brown (Figures 7e and 10c). Laminated pendants can reflect climate variations (Manafi and Poch, 2012). According to the authors, lighter colored blades, composed of relatively pure calcium carbonate, indicate dry periods, not suitable for biological activity. However, in more humid periods, climatic conditions become suitable for biological activity and the development of vegetation, increasing the production of organic matter, mixed with the calcium carbonate, resulting in darker stained  blades (Manafi and Poch, 2012). In paleosols of the Marília Formation, the multilayer laminate of calcite pendant (Pt), indicated by yellow arrows (Figure 10c), point to improvement in moisture conditions. From this tone of brown, it may be inferred that climate conditions were more suitable for biological activity and development of vegetation, with production of organic matter, or the interpretation that the profile was under hydromorphic conditions.
In the paleosols of the Marília Formation, the amount of pending calcite increased in the Btkm horizons in relation to the Bkm of profiles 8 and 9.
Infillings and coatings were common features in calcretes (Figures 7e and 8). Coatings, hypocoatings, quasicoatings, and infillings are practically the result of pedological formation, the production of weathering (Stoops, 2008). There is partial alteration and replacement of quartz by microsparitic calcite coating (calcans), indicated by the yellow arrow (Figure 7e). Coating of quartz by carbonate is a typical feature of soil profiles (Bedelean, 2004). According to Stoops (2008), the juxtaposition of (hypo) coatings allows chronology to be established, while analysis of the combination of fabric to determine the constituents enables micromorphology to identify the origin of the parent material or distinguish if the clay accumulations were formed by weathering or illuviation. The relative chronology of the paleosols of the Marília Formation (Figure 8e) suggested that there was initially carbonate coating, and then the clay was coated with iron oxide, a superimposing process (second stage). The chronological interpretation of  infilling also revealed that palygorskite (P) precipitated in the voids of paleosols, in the case of a secondary mineral (Figure 8a).
Carbonate coatings can be formed by evaporation or mechanical translocation or have their origins related to biological activity, consisting of microbial tubules or needle-like The alkaline pH of these horizons causes a destabilization of quartz and its replacement by calcite, due to the opposite tendency in the solubility of silicon and carbonates as a function of the pH. In this replacement process called brecciation (indicated by the red arrow), the mineral grains are disintegrated into a series of separate fragments by the action of carbonates, but generally preserve the optical orientation of the original crystals. The quartz in the process of weathering presents a parallel linear and dotted alteration patterns with degree 3 (indicating that 75 to 97.5 % of the mineral was weathered); (b) Idem with XPL, a superimposition process of clayey material (iron oxides) on the carbonate features can be perceived with polarized light; (c) Partial alteration process with carbonate replacement in feldspar (Bk/Ck horizon -P8) (yellow arrow); feldspar shows irregular linear alteration pattern with degree 1 (indicating that 2.5 to 25 % of mineral was weathered); (d) Idem with XPL; (e) Parallel linear alteration pattern with degree 3 in biotite (B) of the Bkm1 horizon (P8) (yellow arrow); the red arrow indicates the dense discontinuous infilling feature (pedotubule), represented by the palygorskite mineral (P); Feldspar (F) is also present in the coarse material of the groundmass, as well as quartz (Q) in high proportions; and (f) Idem with XPL. Figures obtained by an optical microscope with a 10X objective and 100X eyepiece.
calcite fibers (Stoops and Schaefer, 2010). These types of carbonate coatings were found in the paleosols of the Marília Formation (Figure 7). The red arrow indicated a product of carbonate reprecipitation around the quartz grain, which was a calcite coating (calcan) in paleosols of the Bkm horizon of the Marília Formation (Figure 7a).
Cytomorphic calcite was identified in profiles 8 and 9. This form of calcite develops under high biological activity and the presence of moisture (Manafi and Poch, 2012).
The pedofeatures of carbonate depletion and accumulations of cytomorphic calcite are the result of biological activity in a wet climate, and are relics of a humid climate in the past (Manafi and Poch, 2012). Cytomorphic calcite composed of a sparitic carbonate decalcification zone found in the Btkm horizons of profile 9 (Figure 10e) can indicate transition from a drier climate to a more humid one, improvement in regional or local hydrological conditions, or that the profile was under hydromorphic conditions.  However, calcification-decalcification pedofeatures associated with impregnating the root tissue is often observed in semi-arid calcareous soils (Durand et al., 2010). This feature, according to the authors, corresponds to infillings of channels by cytomorphic sparitic calcite, surrounded by a zone of non-calcareous material.
Another pedofeature found in Bt horizons was pisolite (Figure 10a). Pisolites are common in highly developed petrocalcic horizons or calcretes (Phases V and VI of Machette (1985) when the calcic horizon is hardened and subject to pedogenesis and alteration by physical and chemical weathering (Durand et al., 2010). Concentric bands of clay in pisolites may result from clay neoformation with Si and Al, and these concentric layers may indicate control by microorganisms associated with roots (Durand et al., 2010).
In paleosols of the Marília Formation, the carbonate character was confirmed by the predominance of porphyric-related distribution and crystallitic b-fabric in the  groundmass ( Figure 4a). Crystallitic b-fabric can develop when fine calcite crystals gradually precipitate in the micromass of clay and when the pore space between the grains is progressively filled by pedogenic carbonate crystals with a particle size between clay and silt (Stoops and Schaefer, 2010). Identification and interpretation of porphyric-chitonic related distribution, speckled-porostriated b-fabric, and clay coatings have shown the features of argillic horizons in the paleosols with a Bt horizon defined in the field (Figures 4c, 4e, 8c, and 8e).
Gefuric-related distributions identified in paleosols indicated aggradation (training) for illuviation and enaulics, an increased pedogenesis process, and stability of the environment. Chitonic-related distributions are results of a poor illuviation process in  The evolution of pedogenic processes, in light of micromorphological analysis, evidenced three moments or stages to the genesis of paleosols. In the drier first stage, there is intense cementation of horizons with micritic and microsparitic calcite (generating porphyric c/f-related distribution) and palygorskite precipitation in the voids or filling bioturbations (Figures 4a, 6a, 6e, 7e, 8a, and 9). The second phase is characterized by the destruction of porphyric-related distributions for establishing chitonic and enaulic-related distributions (Figures 4e and 8e). Relative chronology revealed a superimposition of the features of iron oxides on carbonate cementation (Figure 8e). In the third local or regional wetter phase, carbonate cementation no longer exists, giving way to advancement of chitonic-and gefuric-related distributions, resulting in paleosols with Bt (Figures 4c, 8c, and 10a).
In the Maastrichtian of the Bauru Basin, the interpretation of micromorphology showed that paleosols with a Bkm horizon are older and more arid environments. After that, the paleosols are coated by clay minerals with iron oxides, culminating in the destruction of porphyric-related distributions (carbonate cementation) and the establishment of chitonic-related distributions in wetter local or regional climatic conditions, generating paleosols with a Bt horizon. However, these climate conditions, although more hot and humid, were not enough to generate enaulic-related distributions, typical of current Oxisols.
The genesis of paleosols with Bt horizons predates the establishment of higher humidity and temperature conditions that were established in the Tertiary Period. The paleosols with Bkm were formed first, under drier conditions. The second stage was characterized by destruction of porphyric-related distributions and formation of chitonic-and gefuric-related distributions.
The third stage was probably wetter and there was advancement of chitonic-and gefuric-related distributions, and formation of paleosols with Bt.
These results suggest climatic cyclicality or changes in the hydrology of the basin during the Maastrichtian.